7、Apache Spark与Databricks

SparkSummitJune2014ApacheSparkandDatabricksAdoptionAllmajorHadoopdistributionsincludeSparkBeyondHadoopPartnershipsPartnerwithSparkdistributorstoprovidegreatexperiencetoeverySparkuserPartnersCertificationBuildastrongapplicationecosystemSparkAPISparkDistros…DistrosCertSparkApps…AppCertCertificationFreecertificationprocessScriptsforcertifyingSparkdistributions• Developedbycommunity• Open-sourceAnyonewillbeabletocertifyanySparkdistributionTrainingWe’vebeenteachingSparksince2012• 400+peoplethisyearthroughDatabricksJustlaunchedanewtrainingprogram• Alreadyholdworkshopsin5cities300+peoplesignedupfortrainingonWednesdaySolveBigDataChallengesBigPromiseGreatsuccessesusingBigDataBigPromiseYourcompanyhere!EveryorganizationcollectsdataGreatsuccessesusingBigDataBigChallengeGreatsuccessesusingBigDataYourcompanyhere!Google,Facebookspendbillions$todevelop,implement,andrundataanalysistoolsandproductsEveryorganizationcollectsdataTypicalStoryYourcompanystartsaBigDatainitiativeYouaretaskedto…1)BuildaHadoopcluster2)Buildadatapipeline3)Getinsights&builddataproductsClustershardtosetupandmanageNeedtointegrateazoooftoolsToolsarehardtouse(IT)(engineers,datascientists)(engineers,datascientists,analysts)TypicalDataPipelineDataETLExplorationDashboards&ReportsDataProductsIntegratedisparate,clunkytoolsHardtonavigatedata,developanddeployappsAdvancedAnalyticsVisionMakebigdataeasyFromChallengestoSolutionsChallengesSolutionsApacheSparkHostedplatformInteractiveWorkspaceToolsarehardtouseClustershardtosetupandmanageNeedtointegrateazoooftoolsDatabricksCloudDatabricksCloudDatabricksWorkspaceDatabricksPlatformDatabricksPlatform……DatabricksWorkspaceDatabricksPlatformDatabricksPlatformStartclustersinsecondsZero-costmanagementDynamicallyscaleup&downApacheSparkUnifies• Streaming• SQL• Machinelearning• GraphsSinglesystem,singleAPIDatabricksPlatformDatabricksWorkspaceDatabricksWorkspaceDashboardsNotebooksJobsAppsDatabricksPlatformDatabricksWorkspaceNotebooksSupportPython,SQL,ScalaInteractivecommands&plotsOn-linecollaborationDashboardsWYSIWYGbuilderInteractiveplotsOne-clickpublishingJobLauncherRunarbitrarySparkjobs,programmaticallyDramaticallySimplifyDataPipelineDataETLExplorationAdvancedAnalyticsDashboards&ReportsDataProductsCloudDramaticallySimplifyDataPipelineDataFreeuserstofocusonfindinganswers&buildingproductsETLExplorationAdvancedAnalyticsDashboards&ReportsDataProductsCloudDemoAvailabilityStartedclosedbetaprogramearlierthisyearLimitedavailabilitysoon• Graduallyrampingup• Signupondatabricks.com!3rdPartyAppsDatabricksPlatformDatabricksWorkspace3rdPartyAppsDatabricksPlatform…DatabricksWorkspaceAppsDatabricksCloudandSparkDatabricksCloudruns100%ApacheSpark• Nolockin:anyDatabricksCloudapprunsonanycertifiedSparkdistributionDatabricksCloudacceleratesSparkadoption• ProvideeasiestwaytolearnanduseApacheSparkDatabricksCloudDatabricksPlatformDatabricksWorkspaceMakebigdataeasyDramaticallysimplify• analyzingbigdata• buildingdataproductsFuelgrowthofSparkecosystemThankYou!